Qwen3.5-4B-AutoRound-NVFP4-RTN

Model Details

This model is a NVFP4 (NVIDIA FP4) quantization of Qwen/Qwen3.5-4B generated by AutoRound. Please follow the license of the original model.

Quantization Details

Attribute	Value
Base Model	Qwen/Qwen3.5-4B
Quantization Tool	AutoRound
Quantization Scheme	NVFP4
Quantized Size	4398 MB

Evaluation Results

Task	Accuracy
hellaswag	0.5306
mmlu	0.7171
mmlu_abstract_algebra	0.5000
mmlu_anatomy	0.7259
mmlu_astronomy	0.8421
mmlu_business_ethics	0.7500
mmlu_clinical_knowledge	0.7774
mmlu_college_biology	0.8542
mmlu_college_chemistry	0.5600
mmlu_college_computer_science	0.6900
mmlu_college_mathematics	0.5900
mmlu_college_medicine	0.6936
mmlu_college_physics	0.5294
mmlu_computer_security	0.8000
mmlu_conceptual_physics	0.8340
mmlu_econometrics	0.6842
mmlu_electrical_engineering	0.7655
mmlu_elementary_mathematics	0.6825
mmlu_formal_logic	0.6190
mmlu_global_facts	0.3300
mmlu_high_school_biology	0.8871
mmlu_high_school_chemistry	0.7635
mmlu_high_school_computer_science	0.7600
mmlu_high_school_european_history	0.8364
mmlu_high_school_geography	0.8535
mmlu_high_school_government_and_politics	0.9016
mmlu_high_school_macroeconomics	0.7795
mmlu_high_school_mathematics	0.4704
mmlu_high_school_microeconomics	0.8824
mmlu_high_school_physics	0.6424
mmlu_high_school_psychology	0.9046
mmlu_high_school_statistics	0.6806
mmlu_high_school_us_history	0.8382
mmlu_high_school_world_history	0.8523
mmlu_human_aging	0.7130
mmlu_human_sexuality	0.8397
mmlu_humanities	0.6346
mmlu_international_law	0.8512
mmlu_jurisprudence	0.7870
mmlu_logical_fallacies	0.7853
mmlu_machine_learning	0.5804
mmlu_management	0.8932
mmlu_marketing	0.9316
mmlu_medical_genetics	0.8400
mmlu_miscellaneous	0.8289
mmlu_moral_disputes	0.7572
mmlu_moral_scenarios	0.4279
mmlu_nutrition	0.7810
mmlu_other	0.7531
mmlu_philosophy	0.7460
mmlu_prehistory	0.7593
mmlu_professional_accounting	0.5638
mmlu_professional_law	0.5319
mmlu_professional_medicine	0.7868
mmlu_professional_psychology	0.7582
mmlu_public_relations	0.7182
mmlu_security_studies	0.7388
mmlu_social_sciences	0.8187
mmlu_sociology	0.8557
mmlu_stem	0.7054
mmlu_us_foreign_policy	0.8500
mmlu_virology	0.5542
mmlu_world_religions	0.8304
piqa	0.7606

How to Use

HF Usage

Step 1: Install AutoRound

pip install auto-round

Step 2: Load and run the quantized model

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "Qwen3.5-4B-AutoRound-NVFP4-RTN"

# load the tokenizer and the model
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype="auto", device_map="auto")

# prepare the model input
prompt = "Write a quick sort algorithm."
messages = [{"role": "user", "content": prompt}]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

# conduct text completion
generated_ids = model.generate(**model_inputs, max_new_tokens=512)
output_ids = generated_ids[0][len(model_inputs.input_ids[0]) :].tolist()

content = tokenizer.decode(output_ids, skip_special_tokens=True)
print("content:", content)

VLLM Usage

vllm serve Qwen3.5-4B-AutoRound-NVFP4-RTN \
    --trust-remote-code \
    --dtype bfloat16 \
    --tensor_parallel_size 1

If you encounter any issues, feel free to open an issue on the AutoRound GitHub repo or provide feedback on the Low-Bit Open LLM Leaderboard.

Ethical Considerations and Limitations

The model can produce factually incorrect output, and should not be relied on to produce factually accurate information. Because of the limitations of the pretrained model and the finetuning datasets, it is possible that this model could generate lewd, biased or otherwise offensive outputs. Therefore, before deploying any applications of the model, developers should perform safety testing.

Caveats and Recommendations

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Here are a couple of useful links to learn more about Intel's AI software:

Disclaimer

The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. Please consult an attorney before using this model for commercial purposes.

Cite

@article{cheng2023optimize,
  title={Optimize weight rounding via signed gradient descent for the quantization of llms},
  author={Cheng, Wenhua and Zhang, Weiwei and Shen, Haihao and Cai, Yiyang and He, Xin and Lv, Kaokao and Liu, Yi},
  journal={arXiv preprint arXiv:2309.05516},
  year={2023}
}

arxiv github

This model is part of the Intel Low-Bit Open LLM Leaderboard initiative.